Volume 17, Issue 7, Pages 552-563 (July 2015) Mesenchymal Stem Cells Shed Amphiregulin at the Surface of Lung Carcinoma Cells in a Juxtacrine Manner  Oriane Carnet, Julie Lecomte, Anne Masset, Irina Primac, Tania Durré, Ludovic Maertens, Benoit Detry, Silvia Blacher, Christine Gilles, Christel Péqueux, Jenny Paupert, Jean-Michel Foidart, Guy Jerusalem, Didier Cataldo, Agnès Noel  Neoplasia  Volume 17, Issue 7, Pages 552-563 (July 2015) DOI: 10.1016/j.neo.2015.07.002 Copyright © 2015 The Authors Terms and Conditions

Figure 1 MSCs enhance tumor growth and metastasis formation. Luc-LLC cells (1 × 105) were subcutaneously injected into C57BL/6J mice in both flanks in the absence (LLC) or presence of BM-MSCs (5 × 105) (LLC+MSC). Representative data from three independent experiments are shown. (A) In vivo bioluminescent signal of the primary tumors at day 14 post cell injection (n = 17 per group). (B) Quantification of the luminescent signal as a function of time after cell injection (n = 17 per group; linear regression ***P < .001). (C) Weights of primary tumors resected at day 14 (n = 17 per group; Student’s t test ***P < .001). (D) In vivo bioluminescent signal of lung metastases at day 35. (E) Representative ex vivo bioluminescent signal of lungs resected from mice at day 35. (F) Statistical analysis of metastasis incidence at day 35 (chi-square test **P < .01). Neoplasia 2015 17, 552-563DOI: (10.1016/j.neo.2015.07.002) Copyright © 2015 The Authors Terms and Conditions

Figure 2 GFP-labeled BM-MSCs co-transplanted with tumor cells do not affect the vascular density within a tumor mass. Luc-labeled LLC cells (1 × 105) were subcutaneously injected in C57BL/6J mice in the absence (LLC+; MSC−) or presence of GFP-expressing BM-MSCs (5 × 105) (LLC+; MSC+). Primary tumors were resected at 7, 9, 12, and 14 days (D7-D14) immediately after mice injection with dextran-FITC. (A) Immunodetection of GFP in BM-MSCs (brown) and FITC (vessels in blue) on paraffin sections of tumor tissues. (B) Quantitative analysis of the vessel density in tumors. Data are expressed as the fold induction of the functional blood vessel (BV) number per mm2 that was observed within the primary tumor mass. (C) Measurement of the hemoglobin content in tumors. The amount of hemoglobin was normalized to the weight of the lyophilized tumor (n = 10 per group, ns = no statistical difference). Neoplasia 2015 17, 552-563DOI: (10.1016/j.neo.2015.07.002) Copyright © 2015 The Authors Terms and Conditions

Figure 3 BM-MSCs promote LLC cell proliferation and invasion. LLC cells were cultured alone (LLC, monoculture) or co-cultured with MSCs (LLC+MSC) in direct contact (direct co-culture) or separated by a semipermeable membrane in a transwell chamber (LLC/MSC, indirect co-culture). Cells were either seeded on plastic (A, B) or embedded as spheroids in a collagen gel (C). (B) Quantitative analysis of the luminescence intensity of Luc-LLC cells cultured for 3 days with or without BM-MSCs (Mann-Whitney: **P < .01; ***P < .001). (C) The invasion of LLC cells in the 3D spheroid model is increased in the presence of BM-MSCs (LLC+MSC) (Mann-Whitney: ***P < .001). In the lower panel, the spheroid is composed of LLC cells and GFP+ BM-MSCs, revealing that only cancer cells migrate into the matrix, whereas BM-MSCs remain in the spheroid (green cells). The graph corresponds to the quantification of cell invasion. Neoplasia 2015 17, 552-563DOI: (10.1016/j.neo.2015.07.002) Copyright © 2015 The Authors Terms and Conditions

Figure 4 BM-MSCs enhance AREG secretion from LLC cells. LLC cells and BM-MSCs were cultured for 3 days, either alone (monoculture), together (direct co-culture), or separated by a semipermeable membrane to avoid cell–cell contacts (indirect co-culture), as schematically presented (A). The results presented are those of one representative assay out of at least three experiments using different primary BM-MSC cultures. (B) The level of AREG in conditioned media is compared using a cytokine array. (C) RT-PCR analysis of AREG mRNA expression. The results are expressed in arbitrary units corresponding to the AREG/28S ratio (Mann-Whitney: **P < .01). (D) Western blot analysis of the AREG production in culture lysates. Production of the actin protein is included as a loading control. The graph corresponds to the quantification of AREG production by scanning densitometry. The results are expressed as the AREG/actin ratio. (E) Secreted levels of AREG analyzed by ELISA (right panel) and by Western blotting (left panel) on conditioned medium from LLC cells, BM-MSCs, and direct or indirect co-cultures of LLC cells and BM-MSCs. (F) Western blot analysis of tumors resected 9 or 12 days after injection of LLC cells with or without BM-MSCs (Mann-Whitney: *P < .05, **P < .01). Actin is used as a loading control, and data are presented as the AREG/actin ratio. Neoplasia 2015 17, 552-563DOI: (10.1016/j.neo.2015.07.002) Copyright © 2015 The Authors Terms and Conditions

Figure 5 AREG present in co-cultures of MSC+LLC-conditioned media enhances tumor cell invasion in a 3D spheroid model. LLC cells were cultured as monospheroids (LLC) without or with medium conditioned by co-cultured cells (CM LLC+MSC) or co-cultured with BM-MSC in heterospheroids (LLC+MSC). (A) Increasing concentrations of recombinant AREG (recAREG) were added to monospheroid cultures. (B) A control antibody (Ab Ctrl) or an anti-mouse AREG-blocking antibody (Ab AREG) was preincubated overnight at 4°C with the conditioned medium. The stimulation of LLC cell invasion by co-culture–conditioned medium (CM LLC+MSC) is abrogated upon a preincubation with an AREG-blocking antibody. (C) Monospheroids (LLC) and heterospheroids (LLC+MSC) were incubated with an inhibitor of EGFR (AG1478, 20 μM) or an inhibitor of TACE (TAPI-0, 10 μM). The graph corresponds to the quantification of cell invasion (Student’s t test: *P < .05, **P < .01; ***P < .001). Neoplasia 2015 17, 552-563DOI: (10.1016/j.neo.2015.07.002) Copyright © 2015 The Authors Terms and Conditions

Figure 6 Active TACE is increased in the presence of BM-MSCs and is responsible for AREG shedding. (A) RT-PCR analysis of TACE expression. The results are expressed in arbitrary units corresponding to the ratio between ADAM17/28S. (B, C) Western blot analysis of pro- and active-TACE production in vitro in culture lysates (B) and in vivo in tumors induced by LLC cells with or without BM-MSCs (C). Actin is included as a loading control. The graph corresponds to the quantification of TACE production by scanning densitometry. The ratio of pro-TACE/actin and active TACE/actin reveals a significant increase in active-TACE production when BM-MSCs are present in tumors at days 9 (P = .0002) and 12 (P = .104) after injection. LLC cells and/or BM-MSCs were cultured alone or in co-cultures treated with TAPI-0 (1 μM) or control. BM-MSCs transfected with anti-TACE or control GapmeR were also used. (D) TACE activity was evaluated by FRET at different time points (0 to 2 hours) (linear regression: ***P < .001). (E) AREG amounts were determined in the conditioned medium using ELISA (Mann-Whitney: *P < .05). Neoplasia 2015 17, 552-563DOI: (10.1016/j.neo.2015.07.002) Copyright © 2015 The Authors Terms and Conditions